Novartis: Medicine for the People

Former Harvard professor Mark Fishman bridges the gap between industry and academia at drug giant Novartis, focusing on "unmet medical needs"

In 2003, Swiss pharmaceutical firm Novartis (NVS) tapped renowned cardiologist Mark Fishman to head its new $4 billion global research center in Cambridge, Mass. The former professor of medicine at Harvard University and chief of cardiology at Massachusetts General Hospital took on a make-or-break challenge: to completely redesign the way Novartis develops new medicines. The appointment of an academic with no previous industry experience caught many observers by surprise.

But, three years on, it's clear Fishman and the 1,000 scientists at the Novartis Institutes of Biomedical Research are leading the company to a new period of innovation. Today, Novartis has one of the most promising pipelines in the industry, with 75 potential new medicines in development. On Feb. 21, Fishman, 55, met with London Correspondent Kerry Capell to discuss the changes under way.

What are some of the first changes you made?

First, we established a new headquarters in Cambridge, Mass., where we have hired more than 1,000 scientists, and started some major new initiatives in fundamental and disease-related sciences.

Second, we have better incorporated medicine into all of our research programs, and eliminated traditional borders to clinical development, primarily by hiring of a cadre of physician-scientists. Now, from the start of a project, we decide which patients would most benefit from new medicines, and how trials might be run quickly and safely in them.

Third, we have changed the way success is measured. Companies love to measure progress in quantitative ways, but that is not always sensible for science, and we have instituted metrics that allow for longer-term programs. We are trying to reinforce true innovation. Success is not numbers. Success is new medicines that can help patients. Fortunately, many forces now conspire to reinforce innovation as a driving force. Whereas, in the past, second or third drugs of similar action could reap financial rewards, today regulators, patients, health insurers, and governments are demanding better, safer, and more effective new medicines. So all the stars are aligned to encourage innovation.

Traditionally, big drug companies have focused their research on the biggest potential markets. Their approach has been to focus on single proteins as potential targets for new drugs. How have you changed the way Novartis approaches drug discovery?

We focus on unmet medical needs, meaning there are patients, not markets, needing our medicines. Simply put, medicines affect proteins that cause disease. These proteins are termed drug targets. Today we have a complete list of proteins in the body through the genome project, but which to pick as targets is the toughest part of the discovery process, because our knowledge of how they fit into medical science is in its infancy.

You've said drug discovery needs a "new grammar." What do you mean by that?

The genome to date is a like a list of all the words in the biological dictionary. We know the sequence of 22,000 or so genes, and this sequence predicts the proteins the genes encode. These are the words. But they lack definition. What do all these genes and proteins do, in health and in disease? Are these words reproducibly assembled in predictable arrays, or are they used, 22,000 by 22,000, in totally different ways in different cells at different times?

Fortunately, in many cases, the same sets of proteins interact in similar ways. These assemblages are called pathways. This is the "grammar" of biology. There are far fewer pathways than genes, probably in the hundreds. Key nodes in such pathways may make excellent drug targets.

What are some examples?

There is a pathway responsible for cell growth, present in most cells. Cell growth is key to cancer, and also to the rapid occlusion of vessels that occurs after vessel expansion by stents (termed "re-stenosis" after angioplasty), and to rejection of tissues after transplantation. A drug that targets a key node in this pathway, termed mTor, is under investigation for use in all these apparently unrelated diseases. We are now examining such medicines in rare cancers that seem completely dependent upon anomalous activation of the mTor system.

Are there other examples where Novartis is looking at relatively small patient groups as a means of discovering new medicines for more common diseases?

One disease I am very intrigued by is cystic fibrosis, which is an inherited disease due to a genetic mutation. It is an example of where you can understand the genetics, but still not have an obvious cure. It occurs in children and often leads to frequent hospitalization.

We have been looking at cystic fibrosis, which in some ways resembles certain types of bronchitis that adults get. So we are interested in whether we can go to a group of cystic fibrosis patients, a group which might be more homogeneous and also is in desperate need of better medicines, in order to find effective therapies and, if it works, then extrapolate to a subset of adult patients with chronic bronchitis. We are at early stages of this now.

Two years ago, Novartis Institutes of Biomedical Research announced it was funding an unusual collaboration, with the Broad Institute of MIT and Harvard, to decipher the genetic causes of type 2 diabetes and make the findings freely available over the Internet. How is it going?

It has worked really well for both us and for our partners. As you know, we elected to do this work in the public domain. We are looking for the genes that predispose people to type 2 diabetes and to diabetic complications. There already is evidence that regions of the genome are associated with diabetes, but the specific genes have, with rare exceptions, been elusive. Data from the genetic scans done at the Broad should begin to be available to all by the late spring.

There are other benefits to both parties. For Novartis Institute scientists, this collaboration brings continuous updating on new genomic technologies. For scientists of the Broad Institute, MIT, and Harvard, there is the opportunity to get a lot of insight into how the genome could be used for drug discovery. It generated immense excitement on the part of scientists involved in the project, some of whom have come to work at Novartis.

Why not do the research internally at Novartis?

At the end of the day, we are in the business of trying to discover medicines. The input into that system, the discovery of proteins related to the disease, is really tough, and it is debatable as to how much of that information is patentable anyway. So our view was, the more we could get this out, and the more quickly people could work on it from around the world, the more likely targets would be discovered.

We could have built it up ourselves, as we have a lot of talent internally in genetics. But my feeling was that we also would benefit from the sociology of working in an open way with the academic world. If you want to change any sociology, such as that for drug discovery, it is most efficient to "contaminate" it with a different way of thinking. I wanted to make scientists at Novartis even more critical, open, and entrepreneurial, and lecturing does not suffice. Instead, you have to show them how it works in practice.

We have other open types of collaboration with MIT and Harvard, and we try to structure them in ways so the scientists retain all of their academic rights, and can publish and exchange information. But if something is discovered that is of possible utility to discovery at the Novartis Institutes, we will encourage them to get involved in trying to foster it. So the goal is to share the risk and share the benefit.

So, rather than give them huge amounts of money up front, we give them enough to do their part of the project, and we'll often do part of it ourselves. It is not different from the best biotech collaborations we do, but it is a little unusual in terms of engaging the scientists in the drug-discovery process.

Before it's here, it's on the Bloomberg Terminal. LEARN MORE